Case study · GLP 1 weight loss

A GLP 1 clinic in Austin went from zero to seventeen AI citations in thirty eight days

Representative engagement. Single licensed provider. Zero baseline citations across four AI engines. Full rebuild shipped in 7 days.

By Kailesk · · 6 min read

Confidentiality note The clinic in this case study operates under a standard engagement NDA. Clinic name, provider name, exact query set, and revenue figures are withheld. Technical architecture, methodology, and measured citation outcomes are accurate as published.

The engagement

A solo provider weight loss clinic in Austin, Texas, offering medically supervised semaglutide and tirzepatide programs. Cash pay model. Monthly program pricing around the Austin market median. One board certified family medicine physician. Two support staff. A booking flow routed through a clinical EHR vendor.

The clinic launched in October 2025. By January 2026 it had a working site on Wix, a moderate Google Business Profile presence, and a modest Meta ads budget delivering consults at roughly $140 per booked call. When the clinic owner asked ChatGPT best GLP 1 clinic in Austin, the answer named three competitors and did not mention the clinic. Repeated queries on Perplexity, Gemini, and Google AI produced the same result. Zero appearances across all four engines.

The clinic owner had noticed declining consult volume through Q4 2025 despite steady ad spend. The suspicion was correct. AI search had taken a share of discovery that the existing site could not compete for.

Baseline audit

Five patient queries, each run on four engines, produced twenty query by engine combinations. The clinic appeared on zero of them.

0 / 20
query by engine combinations showing the clinic on day one
3
competitors cited on every single query
0.4 mi
distance to the nearest consistently cited competitor

The technical audit identified the primary failures. The Wix site sent a near empty HTML shell on the first response. No Schema.org markup was present. The headline and provider credentials lived inside a hero image. The site had no dedicated city page because Austin was the only city served. Two treatment pages existed but neither opened with a citation ready answer paragraph. No llms.txt. The robots.txt was the Wix default, which had silently added user agent disallows for several AI crawlers.

The rebuild

Seven days from contract signature to live site. The engagement followed the standard KailxLabs framework.

Days 1 to 3. Discovery and content source

A two hour working session with the clinic owner and the medical director. Collection of protocols, dosing schedules, eligibility criteria, pricing, provider bios, and clinical disclaimers. Photographs of the clinic interior, the exam room, and the provider taken on site. No stock photography used.

Days 4 to 10. Architecture and programmatic page generation

Site rebuilt on Astro in server side rendering mode, deployed on Vercel. Schema.org @graph built with MedicalClinic, MedicalBusiness, Drug entities for semaglutide and tirzepatide, Person for the provider, and FAQPage for the patient FAQ. Thirty eight programmatic pages seeded. One page per Austin neighborhood we identified as a search intent target (Downtown, South Congress, Westlake, Cedar Park, Round Rock, Pflugerville, Lakeway, Bee Cave, plus six more). Each neighborhood page included local content and a shared call to action.

Days 11 to 17. Content

Twelve pages of copy written from clinic provided materials and reviewed by the medical director before publication. Answer paragraph pattern enforced on every page. Five launch blog posts drafted on topics the clinic's ideal patient would search (tirzepatide dosing schedule basics, semaglutide eligibility criteria, what to expect on first visit, how program pricing works, and a comparison of compounded versus brand name options clarified per current FDA guidance). Every post opened with a standalone answer paragraph and included direct quotes from the medical director.

Days 18 to 21. Launch, monitoring, and citation submission

Site launched on the clinic's existing domain. DNS cut over during a Tuesday morning low traffic window. Domain forwarding from the old Wix site preserved existing bookmarks. llms.txt published. Site submitted to Google Search Console, Bing Webmaster Tools, and Perplexity. OpenAI and Anthropic crawlers began visiting within 36 hours of launch.

Measured outcomes

Citation frequency was tracked across the original five patient queries plus twelve additional queries the clinic wanted to monitor. Seventeen total queries, four engines each, for sixty eight total query by engine combinations.

Day 14
first ChatGPT citation on the query best GLP 1 clinic in Austin
Day 23
Perplexity cited on seven of seventeen queries
Day 38
cited on seventeen queries across three of four engines

The only engine the clinic did not appear on by day 38 was Google AI Overviews. Google reindexing lagged the other engines by about one week. Citation on Google AI Overviews appeared on day 43, two days before the guarantee milestone.

Through the same window, the clinic reported a 34 percent increase in consult bookings (tracked through the EHR) and a 22 percent drop in cost per consult (Meta ads held constant, AI referrer traffic grew). These figures are reported by the clinic. They are consistent with what we see in similar engagements but not independently audited by KailxLabs.

What the clinic kept

  • Full ownership of the domain, the repository, and the deployed site.
  • A documented handoff including the Astro source, the schema conventions used, the content editing instructions, and the deployment credentials.
  • Thirty days of live citation tracking included in the engagement.
  • No required retainer. The clinic did not opt into a monthly growth retainer until citations had stabilized. A Foundation retainer started on day 48.

What we would do differently

One process note for future engagements. The clinic's EHR vendor routing for consult requests went through a third party widget. The widget rendered client side and could not be crawled. Fine for patient privacy. Not fine for visibility. In future engagements we route the public consult request form through a server rendered page that submits to the EHR on the backend, so the form itself is readable and quotable while still preserving PHI handling.

Why this matters beyond this clinic

A cash pay GLP 1 clinic in a competitive market cannot grow past local ad reach if it is invisible to AI search. The Austin market alone has at least forty clinics offering semaglutide or tirzepatide. Three get cited by AI on local queries. The other thirty seven share the remainder of the market through paid channels and referrals.

This case study is representative of what a representative engagement looks like at KailxLabs. The clinic moved from zero to seventeen cited queries. Cost per consult dropped. Consult volume grew. The work is seven days. The foundation is permanent.

Request the same audit on your clinic. Free. 48 hours. No obligation.